AI-written scripts in commercials: the risk of handing your brand's voice to a robot

You ask a chatbot for the script of your brand's next film and, in under a minute, you get an organized text with a beginning, a middle, an end and a tagline. In this 12 article series, here at Mentoniano, we bring together research and cases that help you measure how much that speed can cost your brand's reputation.

In the next articles, we will look with you at market research, academic studies and real cases of commercials made with artificial intelligence. We want to put concrete arguments on your desk so you can decide at which stage the technology comes into the process and who answers for the message.

What is at stake in a commercial

In 30 or 60 seconds, a commercial condenses what your brand wants people to feel and remember. Every choice carries weight: the first image, the line that opens the scene, the silence before the turn, the tagline that closes the film. This short space is where you win the attention of the people watching.

A language model writes from statistical patterns and tends to return the most likely solution for each request. In a film competing for attention with thousands of other ads, this creates a serious problem, because the most likely solution is usually the one the audience has already seen many times.

The audience notices, and that perception weighs on the brand

In December 2024, NielsenIQ released a study with more than 2,000 people who watched AI generated ads, of low and high quality. About 150 participants had their brain activity measured by electroencephalogram, and all of them answered a survey after the screening.

The company itself named the effect a "negative halo": the bad impression left by the ad can spill over into the perception of the brand that signs it. If you invest in media, this finding belongs in the math next to the production cost, because the return on the film depends on what the audience keeps in memory.

Brazilian consumers state their preference

The global AI Monitor 2025 study by Ipsos surveyed 23,216 adults in 30 countries between March and April 2025, a thousand of them in Brazil. Among Brazilians, 62% said they prefer advertising campaigns made by humans and 22% preferred those made by AI, according to a report by Meio & Mensagem.

A preference stated in a survey is a signal that deserves your attention, because it reveals the filter through which the audience receives the message. The people watching your commercial already arrive with that expectation.

When a commercial becomes a talking point for the wrong reasons

In December 2025, McDonald's Netherlands pulled a 45 second Christmas commercial made entirely with AI, a few days after its YouTube debut, following a wave of criticism on social media. In 2024, Coca-Cola recreated its 1995 "Holidays Are Coming" campaign with AI and received comments calling the film "soulless" and "creepy". In the third article of the series, we tell these and other cases in detail.

The tone of the reactions gives the measure of the risk. In December 2025, the Merriam-Webster dictionary chose "slop" as its word of the year and defined it as low quality digital content usually produced in large quantities by means of artificial intelligence. You do not want to see your brand's film associated with that word.

How to use the technology while keeping the brand's voice in house

Artificial intelligence can help at several stages of the work: organizing references, transcribing interviews, testing length variations, adapting subtitles. These are support tasks, and they free up time for what calls for human attention.

Defining what your brand wants to say, choosing the story and polishing every line calls for someone who knows the audience, the product and the craft of writing for the screen. In our view, this is the work that makes a person recognize their own life in a 30 second film.

There are also studies with results more favorable to AI in advertising. One of them, released in Brazil in 2026, appears in the seventh article of the series, with the methodological caveats it deserves.